Deploying AI Vision Kits for Advanced Image Processing and Analytics

By Saumitra Jagdale

Freelance Technology Writer

June 24, 2022


Deploying AI Vision Kits for Advanced Image Processing and Analytics
Image Credits: e-con Systems

One of the key aspects of Artificial Intelligence (AI) and Machine Learning (ML) applications is image processing and computer vision. With the rise in complexity and demand for AI and ML applications, it is important to develop efficient, cutting-edge tools for image processing and analytics of visual data. 


e-con Systems in partnership with Qualcomm has introduced a ready-to-deploy AI Vision kit based on the Qualcomm QCS processor with powerful AI processing capabilities, hence suitable for computationally extensive applications. For supporting the interconnectivity and external components in the ecosystem, Qualcomm vision kit includes an e-con Sony STARVIS IMX415-based 4K MIPI CSI-2 camera module, a Qualcomm® QCS610 SoC-based VVDN SoM, and a carrier board. 

The Sony STARVIS IMX415 super low light sensor is used in the 4K camera module, allowing for incredibly clear pictures even in low light. The Sony STARVIS® IMX415 camera module has a classic range of 115.9° for covering wide range visuals. The camera is capable of working at a frame rate of 30fps while capturing the images in 4K (3840 x 2160) resolution, thus capturing high quality images with precision.  

The carrier board operates on a 12V DC supply and provides Three 4-lane MIPI CSI for camera interfaces, MIPI DSI One 4-lane MIPI DSI display interfaces, a USB3.1, USB2.0(Type C), and a USB 2.0 (Host only) along with a 4-bit Secure digital interface (SDIO). Further, it enhances its capabilities by providing WiFi (2.4GHz/5GHz), Bluetooth, and Ethernet connectivity along with a 50-ohm coaxial connector. 

Image Credits: e-con Systems

The kit is engineered to enable powerful computing for on-device image processing with exceptional power and thermal efficiency. In addition to offering the complete vision kit, e-con Systems will also extend its customization support on the camera, carrier board, software, and enclosure. This enables customers to focus on their end applications while e-con takes full control of the vision kit and solution. 

With this ready-to-deploy AI vision kit and extensive customization services for the camera, BSP, carrier board, software, and enclosure, product developers can reduce their time to market by developing AI-based smart camera systems more efficiently and faster than before. 

Apart from offering full customization support, one of the other key value adds from e-con Systems is the fine tunned Image Signal Processor (ISP). The default Qualcomm ISP has been further to bring out the best-in-class image quality and color reproduction, making it adaptable to a wide range of lighting conditions. 

Image Credits: e-con Systems

Also, the kit comes with a small form factor which makes it perfect for compact embedded systems. Further, it can perform all image processing and analytics indigenously without the connectivity or power of the cloud. 

Features such as small form factor, low power consumption, low light performance, high resolution, and the ability to run image-based machine learning and deep learning models at the edge make qSmartAI80 perfect for applications such as service robots, telemedicine robots, telepresence robots, smart checkouts and carts, smart traffic, dash cameras, and a lot more. 

Apart from the complete vision kit, e-con Systems offers a Yocto-built Linux image with a GStreamer-based camera application for "HDMI streaming," H.264 network streaming over Ethernet and WiFi, snapshot image capture in JPEG and RAW Bayer, H.264 video recording, and cameras feature adjustments such as white balance, gain, and exposure. Engineers may utilize Qualcomm's neural processing SDK for AI, Neural network API (TFLite models), GStreamer plugin (DLC models), and GStreamer TFLite plugin to fully exploit the kit's AI capabilities. After free email registration, software and hardware documentation is available on their official website

The Ready-to-Deploy AI Vision Kit Enhances the Edge Vision Applications

The new qSmartAI80_CUQ610 AI Vision kit by e-con Systems in partnership with Qualcomm offers a complete AI vision and analytics solution to a wide range of Vision applications. The vision kit offers exceptional visual quality using the Sony STARVIS IMX415-based 4K MIPI CSI-2 camera module that is capable of capturing clear images even in low lighting conditions. The kit can even process the images indigenously without the need to connect to any cloud server. Moreover, it is also capable of running ML/DL models for visual data. Hence, the kit is a perfect and reliable solution to a wide range of AI vision applications. 

The demonstration of the kit can be found on the e-con Systems’ YouTube channel


Saumitra Jagdale is a Backend Developer, Freelance Technical Author, Global AI Ambassador (SwissCognitive), Open-source Contributor in Python projects, Leader of Tensorflow Community India, and Passionate AI/ML Enthusiast.

More from Saumitra